It is all this informality and hand-waving around atom:id and atom:updated that is the source of any trouble with using atom:feed as an archive format. Given globally unique ids (atom:id) and version tags (atom:updated) for entries, archiving would be trivial even for "ignoramus apps".